An Avida-ED digital evolution curriculum for undergraduate biology

Introduction to Avida-ED/introductory lesson

Avida-ED was introduced in a lab session via a mini-lecture (approx. 15–20 min) followed by the assignment of the Introductory Lesson and Tutorial in The Avida-ED Lab Book. Students worked together in lab teams, which were chosen at the beginning of the semester using the Team-Maker program included in Purdue University’s CATME package (http://catme.org). The Introductory Lesson and Tutorial was modified from the “Introduction to Digital Evolution Handout & Tutorial” written by Johnson et al. (2009) for the Teach Engineering Curriculum for K12 Teachers (teachengineering.org). Students used this introductory activity to gain familiarity with Avida-ED and the concept of Avidians, the Avida-ED program, and the user interface. They also read the article by Zimmer (2005), which introduced digital evolution research and helped students think about parallels between digital organisms and biological organisms.

Avida-ED exercises 1–3

Exercises 1–3 each were introduced briefly in a lab session (approx. 10 min) and time was built into the lab schedule to allow students to work on these exercises with members of the teaching team present to provide direction and feedback. Students then worked individually outside of class on these three exercises, each of which required students to contribute their data to the pooled class data via Survey Monkey. The course instructor (JS) monitored and summarized the data for presentation and discussion in both the lecture section and the lab sections of the course. These pooled class data provided a powerful visual representation of the processes playing out in the Avida-ED program. Each exercise also had an associated ungraded quiz consisting of a set of free response-type questions pertaining to the exercise. Students were required to submit individual responses to these questions through the Desire2Learn course management system.

Avida-ED exercise 1: understanding the introduction of genetic variations by random mutation

The primary learning goal for students in Exercise 1 was to be able to explain what it means to say that mutations occur at random. Students often carry the mistaken impression that evolution itself is random (Garvin-Doxas and Klymkowsky 2008; Mead and Scott 2010b). Exercise 1 addressed this point by showing that while mutations occur at random, natural selection itself is not random. We also explored the concept of a mutation rate, which provided the opportunity to discuss central tendencies and dispersion about a mean.

Students began Exercise 1 by replicating the “@ancestor” Avidian in the organism viewer with a 10 % mutation rate applied. Students were asked to predict how many mutations they expected to see, and then recorded how many they actually observed. Students compared their initial results with those obtained by a classmate. Students were also asked to explore how their mutations were distributed as a function of position within the Avidian “genome” and to record the locations of each of the mutations that they observed. In this way, we illustrated the point that, in theory, every location in a genome is subject to mutation, even though not all mutations will survive to appear in later generations.

The students universally agreed that they expected to observe five mutations in the 50 positions of the Avidian genome when the organism replicated with a 10 % mutation rate. However, the actual pooled class data shared with the students (Fig. 2a) clearly showed that not everyone observed five mutations in each run, as students often mistakenly expect; individual runs varied anywhere from one to 13 mutations. On the other hand, the mean number of mutations was 4.88, very close to and not significantly different from the expected value of 5.00 (1-sample, 2-tailed t test; t = − 0.5163, df = 80; p = 0.607). (Note: All reported statistical analyses were performed in R version 3.2.3; R Core Team 2013) This allowed students to better see how randomness applies in practice. Students were also able to see visually that the 474 mutations they observed as a class were distributed across the Avidian genome, with each of the 50 positions in the Avidian genome experiencing at least one mutation event (pooled student class data shown in Fig. 2b).

Avida-ED exercise 2: exploring fitness, functions, and selection

The primary learning goal for students in Exercise 2 was to be able to explain what it means to say that a trait has increased in frequency or has gone to fixation in a population. The difficulty of thinking in terms of populations instead of individuals is a recurring impediment to student understanding of evolutionary principles and processes (Bishop and Anderson 1990; Shtulman 2006; Gregory 2009), and one of the strongest features of Avida-ED is its ability to provide a dynamic visual representation of population-level processes. In Exercise 2, an individual Avidian is saved that can perform a particular function (the logic function, NOT). This organism is then separately put into environments either with or without the resource for that function (notose), which allows students to see how selection affects a trait when it provides a reproductive advantage.

Students began the exercise by evolving an Avidian population starting with “@ancestor”, a simple self-replicator ancestor. After sufficient time passed (typically 100–200 updates), students located and isolated an individual Avidian that had evolved the ability to perform the logic function NOT. This individual was used to seed two subsequent independent Avida-ED runs (i.e., it was the ancestor in these runs). In the first of these two subsequent runs, there was no resource in the environment for the ability to perform NOT, which meant that there was no reward for performing NOT. After 2000 updates, students recorded the frequency in the population of individuals able to perform NOT (phenotype frequency). In the second run, the metabolic resource notose was included in the environment, which provided an energy reward to an organism if it performed NOT. Once again, after 2000 updates, students recorded the frequency in the population of individuals able to perform NOT.

Pooled data were collected (Fig. 3) and used in class to test the hypothesis that there was no difference in phenotype frequency (i.e., ability to perform NOT) between runs occurring under the two sets of conditions. Use of the actual class data allowed students to see that, as a whole, when the not phenotype was rewarded in the environment (notose was present), its frequency increased in the population over time. Notably, while most of the students obtained clear results in line with the aggregate data (28/32), some students had no data or ambiguous results. Particularly in these cases, being able to see and discuss the data obtained by peers, and to see the overall trend, was extremely valuable.

Avida-ED exercise 3—exploring mutations and selection: pre-adaptive or post-adaptive?

One enduring misconception students hold is that evolutionary processes are forward looking and that events occur in anticipation of some future need (Mead and Scott 2010a). Thus, the primary learning objective in Exercise 3 is for students to demonstrate that they understand that mutations are random and do not occur simply because they are needed. Avida-ED provides a unique and direct way to test this evolutionary principle.

Students again begin the exercise by seeding a population with the @ancestor, making sure that all rewards are turned off (and thus, that the ability to perform a logic function does not confer a selective advantage). Students then begin their Avida-ED run, and stop the run when the first individual appears in the population that can perform the logic function NOT. Students then record the update at which this event occurred. Next, students repeat the exercise, but now set the initial environmental conditions such that the ability to perform NOT is rewarded (notose is present in the “medium”). Once again, the run is stopped when the first individual appears in the population that can perform NOT, and students record the update at which this event occurred.

When Exercise 3 was introduced with a clicker question, 65 % of the students who responded indicated that they expected the mutation to occur sooner when it was rewarded. However, the actual aggregated data shown in class (Fig. 4) provided students with a different view. Almost all students had the function NOT occur in their populations more quickly under one condition or the other. However, the pooled data showed that there was no difference in the time to first appearance for a mutation in populations with or without reward (2-tailed binomial test, p = 0.541).

Independent research investigation—experimental evolution project with evolving digital organisms

Lark (2014) showed that students do better when they have substantial time and scaffolding to make sense of what they see in Avida-ED. After the students had completed the three preliminary exercises, and had time to explore the data they had generated as a class, the student teams were challenged to come up with a research question of their own to test using Avida-ED. This independent investigation component of the curriculum provided a way for students to do science and think like biologists, and we incorporated into these investigations all of the elements described by Thornton (1972) as adapted by Sundberg and Moncada (1994). Students were told that the purpose of the exercise was to engage in investigation, students consulted with teaching team members to formulate problems and investigatory procedures, students were provided ample time to repeat and/or modify experiments, and students prepared and presented both written and oral reports describing their work.

The Independent Investigations consisted of an initial exercise in which students worked within their research teams to write a research proposal that included a description of their research question, framed the question as a formal hypothesis to be tested, described the experimental design (including descriptions of relevant variables, what data they would collect, and how many replicates they would carry out), and stated a prediction of their expected outcomes if indeed their hypothesis was true. Teams then presented their proposals to the entire class with each team member explaining one part of it. The Guidelines and Expectations for the Proposal are included as Additional file 2, and a sample presentation slide set from one of the research teams is included as Additional file 3.

Student teams carried out their experiments in the teaching lab over the following 3–4 weeks. During this time, the teaching team met regularly with the student teams to check on progress, go over preliminary data, and suggest directions for next steps, much as would happen in an actual research laboratory. Students were also counseled with respect to how to organize their data within spreadsheets, and summarize their data in figures and tables that could be used in their poster presentations. How to organize the data collected during the experiments proved to be one of the biggest challenges faced by the teams, prompting one student to comment that, “I’ve never had so much data before!” A sample Excel file generated by a student team is included as Additional file 4.

Students put together their research posters using a set of guidelines and expectations that were distributed and discussed in the lab (Additional file 5). Preliminary drafts of the posters were discussed during a lab session and formal feedback provided for each group. During the Poster Presentation Session at the BEACON Center, each research team of four students was divided into two teams of two, and each team of two presented their team’s poster and fielded questions about their research. A sample poster from one of the research teams is included as Additional file 6.

The Avida-ED Independent Research Project allowed students to engage in authentic science practices (“do science”), thus addressing a number of our science process skills objectives. Students worked in teams to brainstorm ideas, ask questions, develop hypotheses, design experimental plans, conduct experiments and collect data, analyze data, and write results for presentation to peers and experts. Students were given time to think about their projects and feedback was provided by the teaching team both informally in conversations in the planning stage and formally in response to their proposal presentations. All of these elements provided practice for our students doing the kinds of things that scientists do when they conduct research.

Preliminary assessment of student learning using Avida-ED

In both spring semester 2014 and fall semester 2014, students using Avida-ED were given an assessment item asking them to “Explain how a microbial population evolves resistance to the effects of an antibiotic”. In spring 2014, a quasi-experimental protocol was employed and responses from students in the Avida-ED classroom (n = 100) were compared to responses from students not using Avida-ED (control; n = 87). We collected all data following Michigan State University Institutional Review Board guidelines, IRB #i040365. Preliminary analysis of these data, using content analysis (Weber 1990) to identify emerging themes and the frequency of their occurrence in the student responses, indicated that the students in the Avida-ED classroom more frequently mentioned the key concepts “random mutation” (t test; p < 0.01) and “DNA or gene” (t test; p < 0.05) in response to the prompt than did the students in the classroom that did not use Avida-ED (Fig. 5). The Avida-ED students also used the naïve ideas of intentionality and teleology less often in their responses than did the students in the non-Avida-ED class (t test; p < 0.01).

In fall semester 2014, we employed a pre-/post-test design using the same question for the Avida-ED classroom (n = 39). In this case, the resulting text responses were used as an input file for the EvoGrader application (Moharreri et al. 2014). A higher proportion of the students displayed “Pure Scientific” reasoning in the post-test (Fig. 6a) and students in the post-tests strengthened connections between core concepts and mentioned fewer naïve ideas (Fig. 6b).

Students affective response to the Avida-ED curriculum

Student affective responses to the Avida-ED curricular materials were assessed at the end of the semester in both spring 2014 and fall 2014 using an Avida-ED User’s Survey (Additional file 7). Overall, student response to the Avida-ED implementation was better in fall 2014 (full implementation) than it was in spring 2014 pilot implementation. Evidence in support of this claim was obtained from the item in the User’s Survey asking students which category best describes their overall enjoyment of Avida-ED (Fig. 7). In fall 2014, the percentage of students who “loved” Avida-ED increased by approximately tenfold, while the number who “hated it” decreased to zero. These results are consistent with the findings of Lark (2014), who found a positive correlation between the use of best practices in the implementation of Avida-ED and positive student affective responses.